Payments modernisation is broadening at the same time that banks are being asked to make it commercially useful. Real-time clearing, cloud infrastructure, artificial intelligence (AI), tokenised deposits and new digital asset rails are arriving alongside a fraud environment in which scams increasingly exploit faster and less reversible payments.

At Sibos 2026, ACI Worldwide executives described fraud controls moving into production and modernisation being justified increasingly by growth, while arguing that digital assets should be incorporated into established high-value payment infrastructure rather than supported through separate operating stacks.

Scotty Perkins, Executive Vice President of Product Management at ACI Worldwide, said banks are looking to modernise so they can bring innovation to customers and grow revenue faster. Marc Trepanier, Director of Analytics and Optimisation for Fraud and Financial Crime, said scam detection is shifting towards mule detection and account validation. Together, their responses point to a modernisation agenda in which new capabilities have to become usable parts of the bank’s existing operating environment rather than isolated technology projects.

Scam prevention is moving upstream

Trepanier said scam detection is one of the strongest themes in conversations with financial institutions at Sibos. Fraud teams are moving beyond monitoring outbound debits and trying to identify mule accounts receiving fraudulent funds, while account validation is becoming a basic control across payment rails.

He described account validation as one tool rather than a complete answer to fraud, but said institutions are moving it towards production and that it should work across payment rails. Looking two to three years ahead, he said foundation models for mule detection will become commonplace, while account validation will become a basic fraud risk management control.

AI itself is not new to fraud detection. Trepanier said AI-based fraud scoring has been used for more than two decades. What is changing is where institutions want to apply it: mule detection, model explainability and operational support for fraud teams are moving higher on the agenda.

He stressed that banks need to be able to replay a transaction or event and understand why it received a particular score. He said fully autonomous AI is probably still a few years away. For now, human-assisted AI remains appropriate while institutions build confidence that agents will behave as intended and avoid hallucinations or uncontrolled actions.

Modernisation is being driven by growth, not only cost

Perkins said the imperative to modernise payment infrastructure remains high. He cited Datos Insights research sponsored by ACI, a survey of 200 institutions across North America, Europe and Asia, in which many banks ranked payments modernisation among their top three enterprise initiatives.

The more important finding for Perkins was what institutions said is driving that investment. He said the priority is not principally infrastructure cost savings, but the ability to bring more innovation to more customers and grow revenue through the bank faster.

That changes the business case. Cloud, application programming interfaces and new payment infrastructure are not ends in themselves if they do not shorten the path to new products and services. Perkins positioned ACI’s cloud-native platform strategy as the foundation on which banks can add capabilities including digital assets, tokenised deposits and agentic commerce while retaining the guardrails and resilience required for high-criticality payments.

Tokenised money has to fit the existing payments estate

Perkins said digital assets are a major theme at Sibos, particularly tokenised deposits and their role in settlement. His argument is that banks should not have to build a separate technology estate simply because value is represented on a new ledger.

ACI announced on 22 September that ACI Connetic would support payments orchestrated through Swift’s ledger alongside existing payment flows within a single operating model. The company said the capability would allow banks to process tokenised-deposit payments alongside traditional payment types using the same operational controls, workflows and infrastructure.

That announcement corroborates the operating model Perkins described in the interview. He said institutions should be able to interact with tokenised deposits and digital-asset ledgers as an extension of high-value payments, rather than seeking a separate integration for each new form of money. ACI is also demonstrating how its platform can connect with Swift’s ledger so banks can use that infrastructure without building an ad hoc solution.

Real-time infrastructure has to produce a commercial return

Perkins said banks are becoming more explicit about how modern infrastructure can support monetisable services. As clearing and settlement move from disparate batch processes towards real-time transactions, institutions can rethink when liquidity moves and what services they offer around it.

He gave the example of settlement taking place at a specific point in the day rather than relying on multiple batch cycles. As underlying payment systems allow more real-time clearing and individual real-time gross settlement transactions, he said banks could develop services around those flows and sell them to clients as new products.

Perkins said modern infrastructure can support new services, while Trepanier said AI can help smaller fraud teams strengthen controls.

Fraud controls and digital assets are becoming infrastructure capabilities

Looking over the next two to three years, Trepanier said foundation models for mule detection will become a normal part of fraud infrastructure, while account validation will become a basic fraud risk management control. Perkins similarly said foundation models will extend into areas such as digital identity and automated consumer and commercial payment experiences, provided institutions can unlock the underlying data to make those models more useful.

Perkins also said digital assets are becoming part of financial infrastructure rather than remaining an experimental capability. He said the question of whether institutions will use distributed-ledger technology has effectively been answered; the challenge is enabling it without requiring extensive individual innovation and customisation inside each bank.

For ACI, that brings the themes of the interview together. Banks are trying to put more capabilities into production while avoiding a corresponding multiplication of technology and operating models. If modernisation can let them add fraud controls, real-time services, AI and tokenised settlement on infrastructure that remains resilient and governable, the return is not only avoiding additional complexity but creating a faster route from new technology to products clients will pay for.